A system for the objective characterization by computer of cells in the urinary sediment of man is being developed. The system utilizes image analytical techniques which have provided reliable and highly discriminating information on classification of urothelial cells and on diagnostic patient profiles during the initial period of this grant support. Progress has led to the point where application to actual clinical diagnosis appears promising and feasible. A number of methodological problems must be solved for the development of a practical clinical laboratory system. These are: acquisition of a larger data base to include cell parameters not hitherto included in the computerized cell populations; standardization of sample preparation; development of adaptive cell image processing techniques to control staining variability, and finally, packaging of the software that is being developed in this project so that the analysis can be conducted on an inexpensive laboratory computer. The project explores the potential of currently available computer technology for the cytologic diagnosis of human cancer and provides baseline information required for the automation of cytologic diagnosis of bladder cancer. This proposal is for the renewal of the Grant No. 5-R26-15803-5, currently in force.